Original source of the data is:
https://github.com/wayfair/WANDS
Train, dev, test split of 3:1:1 as per footnote 5 in
https://arxiv.org/abs/2307.0037042,994 candidate products
480 queries
233,448 (query,product) relevance judgements
Product columns:
product_id - ID of a product
product_name - String of product name
product_class - Category which… See the full description on the dataset page:
https://huggingface.co/datasets/napsternxg/wands.
